Linear AI in 2026: What Agent and Loops Actually Do
Linear’s AI is no longer a summarize button. As of 2026 it’s an agent that lives in your workspace, plus Loops, which fire on a schedule or on an event and run work in the background. That last part matters, because most AI assistants still can’t do it.
There is one thing nobody tells you up front: no Linear plan includes AI credits. Not Free, not Business, not Enterprise. Here’s what each piece does, what it actually costs, and the specific kind of work that still falls outside it.
Linear Agent
Per Linear’s docs, Linear Agent “turns your workspace context into action. It lives inside Linear, understands your issues, projects, teams, and history, and can act on your behalf through conversation.”
In practice it creates and updates issues, projects and milestones, summarizes work and customer requests, answers questions about the workspace, and manages comments in threads. You reach it three ways: an agent chat on ⌘/Ctrl + J, mentioning @Linear in a comment, or the dedicated Linear chat. It also answers to @Linear in Slack.
Linear Agent is available by default in your workspace, including on Free. That’s a genuinely generous default, and it’s the piece most teams will use daily.
The Business plan ($16/user/month billed yearly) layers on Triage Intelligence, Code Intelligence, and Linear Asks.
Loops are the interesting part
Loops let you “define scheduled or event-driven operations for Linear Agent to run across your workspace.” When you build one you “choose whether it should run on a schedule or in response to an event.”
That is a real event listener, not a timer with better marketing. A triage automation can tag Linear Agent straight in so it picks up an issue the moment it’s triaged, rather than waiting to be asked. If you’ve watched other AI assistants promise unattended work and deliver a cron job, this is the difference.
Two constraints. Loops are Business and Enterprise only, and they use AI credits to run.
The credit meter, and the part that surprises people
Linear’s AI credits are opt-in and, unusually, not bundled with any plan. Linear states it plainly: “if your workspace never adds funds, you will not have access to these additional features, and you will never be charged for them.”
So a Business seat at $16/user/month buys you the ability to run Loops and Coding Sessions. Running them is a separate purchase.
| What it costs | Rate |
|---|---|
| Loop run (no coding session) | typically $0.07–$0.20 |
| Coding session, model tokens | provider-published rates, no markup |
| Coding session, sandbox runtime | $0.25 per 20-minute block |
Credit that where it’s due: passing model tokens through at cost with no markup is a fairer meter than most of this category runs. Unused balance “rolls over across subscription billing cycles,” independent of your subscription period. The catches are that funds “expire 12 months after the purchase date” and are “non-refundable and non-transferable,” and that when you run dry, “features that require a balance will not be able to begin new work.” Workspaces can drift a few dollars negative if several tasks run at once, then stop.
Price the meter before you make a Loop load-bearing, because the person who discovers the balance hit zero is whoever needed it next.
Where it stops: everything orbits the workspace
This isn’t a knock, it’s just where the gravity sits. Linear Agent reasons over issues, projects, teams and history that are already in Linear. Loops fire on Linear events. Linear also exposes its own MCP server at mcp.linear.app/mcp, so outside clients can read your Linear data.
What none of that does is start from your inbox. Nothing in Linear watches an email arrive from a customer, reads the thread, decides it’s a bug, files the issue, replies to the customer, and puts the follow-up on your calendar. Linear’s event vocabulary is Linear’s own object model, and that’s exactly the right design for an issue tracker.
The problem is that most of the work around an issue tracker doesn’t start in the issue tracker.
Where Carly fits
Carly works the other direction. It’s a full AI executive assistant that connects Linear natively alongside your email, calendar, tasks, Slack and around 260 more, and it triggers on events that happen outside Linear.
- It reaches everything, not a shortlist. Around 260 apps across 45+ categories connect natively, and anything else with a public API connects with your own key from Carly’s integrations. There is effectively nothing in your stack Carly can’t reach, which is the whole point: one assistant instead of an agent per tool.
- The inbound trigger Linear doesn’t have. A Carly workflow can start when an email lands, on a schedule, or on a webhook, then create the Linear issue, draft the reply, and post to Slack as one unattended run.
- The pricing shape is different. Free Zapier-style workflows; AI agents from $35/month. The plumbing that moves data between Linear and everything else isn’t drawing down a credit balance per run.
- The MCP is free. Connect your accounts to Carly and point ChatGPT or Claude at
carlyassistant.com/mcp, and one conversation reaches Linear plus your mail and calendar together.
These two aren’t fighting over the same job. Linear Agent is the best thing available for acting on work that already lives in Linear, and Loops are a real answer to background automation inside it. Carly is for the work that starts in an inbox or a calendar and has to touch four apps before it’s done. Plenty of teams run both, with Linear as the system of record and Carly as the thing feeding it.
More on the connection patterns in best Linear integrations and Linear MCP.
FAQ
What is Linear AI? Two things. Linear Agent is a workspace agent that creates and updates issues, projects and milestones, summarizes work, and answers questions about your workspace through conversation. Loops are recurring agent workflows that run in the background on a schedule or in response to an event.
Is Linear Agent free? Linear Agent is available by default in your workspace, including on the Free plan. Loops and Coding Sessions are the metered pieces, and Loops require the Business plan at $16/user/month billed yearly.
How much do Linear AI credits cost? No plan includes credits. You add funds yourself, and if you never do, you’re never charged and never get access to the credit-gated features. A Loop run without a coding session typically costs $0.07 to $0.20. Coding sessions bill model tokens at provider rates with no markup, plus $0.25 per 20-minute block of sandbox runtime.
Do Linear AI credits expire? Unused balance rolls over across billing cycles, but funds expire 12 months after the purchase date and are non-refundable and non-transferable.
Can Linear AI trigger on an event? Yes, inside Linear. Loops can run in response to a Linear event rather than only on a schedule, which is more than most AI assistants offer. The events are Linear’s own, so nothing fires because an email arrived or a meeting ended. For that, Carly runs the trigger and writes the result into Linear.
Does Linear have an MCP server?
Yes, at mcp.linear.app/mcp, which lets compatible AI clients read and act on your Linear data. Carly hosts one too at carlyassistant.com/mcp, and connecting it costs nothing, so a single ChatGPT or Claude conversation can reach Linear and your mailboxes in the same breath.
More: Best Linear integrations · Linear MCP · Linear alternatives · Linear vs Jira
Ready to automate your busywork?
Carly schedules, researches, and briefs you—so you can focus on what matters.
See what people say
"Before Carly, I relied on a Calendly link, but the whole process felt impersonal and not very professional. Carly changed that by handling all the back-and-forth, so I'm no longer stuck in endless email threads trying to line up schedules.
Now Carly reaches out to candidates, shares my real-time availability, lets them pick a slot, then sends a Zoom link and drops it straight into my calendar. She sends reminders to both of us before each call, which has significantly reduced no-shows and last-minute confusion.
On top of scheduling, Carly acts like a full executive assistant, sending me my schedule the night before so I can prepare for each call. It reminds me of the old x.ai assistant, but Carly is noticeably smarter, faster, and better suited to my healthcare recruitment business."
